Face to Face with Artefacts

militaryzoologistAI and Robotics

Dec 1, 2013 (4 years and 7 months ago)


Face to
ace with

Victor Zwimpfer


This paper aims at conceptualizing interface in terms
of ‹coupling actants›.

This idea

introduced by a theoretical exploration of action

interactivity is considered as the
key eleme
nt to describe the main trait of

the starting point of interface design lies in ‹observing
In contrast to

approaches of action as user
driven and goal
oriented events I propound a

concept of action which makes possible to
the entangled interplay between human beings and artefacts.

By referring to
‹dependency grammar›, ‹structural semantics

, ‹actor
theory›, and
‹agent technology›
I focus on the notion of


and their essential role in
forming actions

y observing actions in this way interface is conceived of as a
spatiotemporal texture which emerges from face to face relations among


nterface, interaction, action, actant, mediation, contingency
, observing


We cannot but f
ace artefacts. From our birth we are embedded in a pre
designed world. However, by acting in it we re
design it all the time. Be they
physical tools, thought devices or even social patterns, artefacts are
unmistakable signs of a particular culture we live
with. In the evolutionary
‹game of life› we have to acknowledge them as players or team
Whether we want it or not artefacts both determine and are determined by the
ways we interface with the environment. Hence human beings and artefacts are
intertwined. The interplay between artefacts and human beings
constitutes the so
. In other words, the scope of action is given
by the interface. But how is this connection between interface and action to be


We often talk of inter
face with reference to screen design or graphical user
interface. However, there are other approaches. Thus interface, according to
this suggestion, could be seen as an overarching label of design, a kind of a
general approach to design.

In this vein inte
rface describes neither an object
nor a subject but a sort of relationship. This notion of interface gives support to
the development of a new model of design. It enables us to bypass an object
centered as well as a subject
centered approach and thus shed
light on the
essential interplay between artefacts and human beings. To put it succinctly,
neither a function nor a user but action itself instantaneously forms the
interface. There is no chief actor who does something on purpose and must be
efficiently su
pported doing this.

Therefore I suggest giving equal status to human beings and artefacts. By
describing action as a temporal association of ‹actants› (both humans and non
humans), we gather a new point of view in interface design. In this sense

is seen as an area in which actants

each other or have

each other. Thus the proposed starting point of interface design lies in
‹observing actions›. To make this approach clear I will put emphasis on the
notion of actants and their const
itutional state of action.


Where ‹actants› are concerned, I think there are at least four approaches which
are worth looking at a little more closely. They will help us to understand the
idea of action which I pursue in this paper. I will
first ju

stress the main idea
of ‹dependency grammar›, ‹structural semantic
›, ‹actor
theory›, and
‹agent technology› which seem important to me in developing an appropriate
concept of action.

Dependency grammar

The term ‹actant› is originally used in
the field of dependency grammar. There
the focus of analysis lies in the dependency structures which are provided by
the verb of a sentence. The unit of investigation is given by the syntax of a
single sentence, which is simply defined as a sequence of wor
ds. The main
decision in this approach is to set the verb as the structural centre of prime
importance. So the syntax order is configured by the part of the sentence which
expresses an action or something which is allowed to happen. Between the
words there

are connexions of dependency. They describe the patterns of
structural relationships among the different word types. Using a stem diagram
one can show the structure of the sentence in a hierarchical order.

Let me give
an example to illustrate this. The s
entence ‹Man acts.› will be simply depicted


See for insta
Simon, Herbert A. (1970): The Sciences of the Artificial, Cambridge, Massachusetts;

Bonsiepe, Gui (1999): Interface. An Approach to Design, Maastricht; Jonas, Wolfgang (2002):
Systemtheorie und Designpraxis, S. 172
188 in: Tom Sommerlatte (Hrsg.), Ange
wandte Systemforschung.
Ein interdisziplinärer Ansatz, Wiesbaden.


For more details see for

Tarvainen, Kalevi (1981): Einführung in die Dependenzgrammatik,
Tübingen, or Weber, Heinz J. (1997): Dependenzgrammatik, Tübingen.


Building hierarchies
is an effective way of structuring and designing organization. Thus it is an ordering
principle (method) of managing complexity.


In such diagrams the verb always serves as the main node which its
completing supplements can be derived from. So to build a grammatical
correct sentence the verb opens a space where related actants take
place. A
verb suggests an action which can only be accomplished through the
connection of involved actants. In the preceding example

is only one
holder or argument place occupied, namely the subject of the sentence.

Nevertheless it obviously s
hows that it is not only the two words but also the
relation between them which constitutes the meaning of the sentence. Thus
mediating relations are essential to the notion of actants. Relations will be
established by so called ‹elective affinities› (Wahl
verwandtschaft) of the verb.
It always owns a valency which must be valued. There are compulsory and
optional place
holders to be instanced. This means that the verb demands a
certain number of actants in order to formulate a grammatically correct
. Transitive verbs for instance are generally followed by an accusative
object whereas intransitive verbs are not.

To summarize the main issue of dependency grammar I would like to stress
the following points:

The verb is the determining structural elemen
t of a sentence which all
other elements

so called actants

are derived from. It implies that the
notion of action is the essential issue.

It is the relation between the actants which constitutes the syntax and
semantic of a sentence. Without any corre
lation between the actants a
specific meaning would not arise.

An appealing and important side effect of such a concept is that actions
are attributed not only to human beings but to every kind of supplemental
position of the verb.

Structural semantic

he approach of structural semantic is based on dependency grammar. But its
field of application is not limited to the sentence as basic unit. Rather it is
extended to the vast realm of narrations. To put it succinctly: stories are made
of actions. Stories
can only be told by assembling actions to a coherent
narrative form. In other words: to answer the question of what is going on we
draw upon the performance and representation of actions.

In this approach
actants serve as the functional description of the

action structure. Not only
characters or human beings are able to act but also animals, objects, things,
concepts, etc. As in the above mentioned approach of dependency grammar


The reader is free to come up with more complex examples. Due to my intention of developing an action
concept based on the no
tion of actants the above mentioned example should be sufficient to get a clear idea
of the principle.


I do not want to dive into the question in what sense the verb ‹to be› would imply any action. However
there must at least be an observer who has to es
tablish a proposition of existence.


Just think of a simple sentence like ‹it is raining cats and dogs› or ‹the car drove into the wall›.


more details see
Greimas, Algirdas Julien (1971): Strukturale Semantik, Braunschweig
, particularly p


his is a familiar idea from the Aristotelian poetic.


the predicate is seen as the nucleus of any given structure of action. «An act
can be thought of as that which accomplishes or undergoes an act,
independently of all other determinations. Thus, to quote L. Tesnière, from
whom this term is borrowed, ‹actants are beings or things that participate in
processes in any form whatsoever
, be it only a walk
on part and in the most
passive way.»

Structural semantics is a way of theorizing about narratives. By observing told
stories and narratives it seeks to find the deep structure of narrations which are
basically driven by actions. This

attempt to reveal invariable patterns and
functions of narrations led to an actantial model. This model offers an
underlying grammatical structure of the basic functions of any action. It is
allegedly applicable as a universal instrument to analyse modes
of actions. The
isolated basic functions that constitute the fundamental elements of all
narratives serve as a kind of typology to describe the development of events.

Thus relations between actants grasp narratives as whole and the actants
indicate the pl
holders that must be occupied when telling a story.

To summarize the main issue of structural semantics I would like to put
emphasis on the following points:

The concept of actants is not limited to the sentence as the basic unit.

Action is seen as th
e crucial element of narratives. It is constituted by
actant relations that produce the development of events.

Actants are any kind of active or passive subjects and objects which piece
together a specific action. The term actant replaces the dichotomy of
subject and object and thus helps to overcome an essential indication of
both subject and object in terms of ontology.

The concept of structural semantics could be seen as too rigorous because
it is based on a fixed deep structure that neglects the evoluti
onary impact
on ‹real life›.



The branch of research of the actor
theory belongs to the field of
science studies. In this field it was concluded that not only humans but also
humans play a key role in laboratory prac
tices. ANT brings into focus the
idea that it is more useful to observe processes than entities of a given world.
By doing this there are no more existing components at stake. Rather
laboratory trials bring forth the actual components called actants. On be
half of
situational action actants emerge and come into existence. The experimental
situation functions as a milieu in which actants will take shape. The main task
is to define actants by observing behaviours and performances occurring in
field studies or
under conditions of experiments.


Greimas, Algirdas Julien; Courtés, Joseph (1982): Semiotics and Language. Analytical Dictionary,



To name the basic functions would not add any value to the further elaboration of th
e actant concept that
is at issue in this paper. Readers interested in this topic are referred to the seminal work of
Algirdas Julien (1971), op. cit.


For more details see
Latour, Bruno (1999): Pandora's Hope. Essays on the Reality of Science St
Cambridge, MA,
particularly p 174ff


ANT takes up a topic well known in linguistics. To replace the questionable
distinction between subjects and objects it refers to the opposition of
syntagmatic and paradigmatic relations of signs. This pair of terms forms
dimensional area often presented as syntagmatic and paradigmatic axes.
The syntagmatic axis determines how signs combine into sequences whereas
the paradigmatic axis determines the semantic value of a sign. In other words:
The syntagmatic dimension d
emonstrates a combination or connection of signs
this…›) and the paradigmatic dimension demonstrates a
selection or substitution of signs (‹this

Both syntagm
and paradigm relations are necessary to understand the m
eaning of a sentence.
The two relations complement each other. «While syntagmatic relations are
possibilities of combination, paradigmatic relations are functional contrasts

they involve
. Temporally, syntagmatic relations refer
ally to other signifiers

within the text, while paradigmatic
relations refer intertextually to signifiers which are

from the text.»

To give an example let us consider the following figure (see next page). In this
figure I have put toget
her some syntagm and paradigm relations that came
spontaneously into my mind when I was thinking about examples in the field
of interface design. In such kinds of variety pools there are multifarious
possibilities of combining the words on the two differen
t axes. For me it was
interesting to see how the meaning would be displaced by changing the
syntagmatic relation through substitution of elements of the paradigmatic axis.
It is up to the reader to figure out which combinations make sense. However, I

like to pursue this topic in another paper as soon as possible because I
think it will give us more insight into the varied notions of the ways we could
understand the phenomenon of interface. At the moment I just want to point
out that the term interface

is placed both as noun and as verb.


In ANT the terms syntagm and paradigm are replaced by association and
substitution respectively. «These two pairs of terms replace the obsolete


Chandler, Daniel (2002): Semiotics. The Basics, London, p 80.


Ibid., p 80 (italics in the original).


distinction between objects and subjects.
In linguistics a syntagm is the set of
words that can be associated in a sentence (‹the fisherman goes fishing with a
basket› thus defines a syntagm), while a paradigm is all the words that can be
substituted in a given position in the sentence (‹the fishe
rman›, ‹the grocer›,
‹the baker› form a paradigm). The linguistic metaphor is generalized to
formulate two basic questions: Association

which actor can be connected
with which other actor? Substitution

which actor can replace which other
actor in a giv
en association?»

Furthermore associations

or as I would like
to say, actions

are in a state of transition more or less all the time. There are
always pre
positions and post
positions which embed the actual situation in a
context. So associations opera
te in both space and time. In a sense they are
events that take place in the present and are related to the past as well as the

ANT offers a further feature which is crucial to the development of an
appropriate action concept. By naming both human

beings and artefacts as
actants this approach gives equal status to them. Human beings do not
dominate artefacts, and neither do artefacts control human beings. Rather, it is
the interplay of all involved actants which rules the situation. Humans and
humans are folded into each other without one part being predominant. It
is the emergent web or network that counts. So the mediation between
technology and man brings forth a new quality which relies on the associated
relationships of actants. In the shap
e of humans and non
humans, actants are
entangled with each other and emerge as hybrid actors. The association is
wholly symmetrical. The following example may give further evidence to the
discussion at issue. Just imagine the relationship between yourself

and a gun.
In what sense does a gun add something to the situation you are in? Does it
just serve as means to an end that you want to achieve? Or does it modify your
goals in terms of the proverb that ‹opportunity makes a thief›? In other words:
Is a gun
simply a neutral device or not? ANT offers the following answer:
«You are different with a gun in your hand; the gun is different with you
holding it. You are another subject because you hold the gun; the gun is
another object because it has entered into a

relationship with you.»

neither you nor the gun are fixed essences but propositions of
. You
are becoming part of a hybrid actor comprising gun and yourself as gunman.

Therefore it is important to have a brief look at the meaning of mediat
between humans and non
humans, between social and technical factors
respectively. Mediation constitutes an event and this generally means that it
cannot be portrayed by causal descriptions. Thus the complications and
entanglements of mundane practices
are not explicable by mere intermediary
links but by an emerging event that is not deducible from the law of causality.

To become familiar with this notion of mediation I will briefly discuss the four
different meanings of technical mediation that are pro
posed by ANT:

Goal Translation
: A program of action is always influenced by other
programs of action. As this happens the interference of the involved
actants' diverse goals results in a translation of the individual goals. The


Latour, Bruno (1999), op. cit., p 303f.


Ibid., p 179.


original goals will be modif
ied by the interplay of the actants. In practical
situations actants are influenced by the co
acting actants and thus the
goals deviate from their original meanings and produce a composite goal.
Thus, the translation of goals suggests that the idea of a hu
man being
perseveringly and rationally aiming at something is rather illusionary.
There are many possibilities of straying and digressing due to the
interference of actants. «I used translation to mean displacement, drift,
invention, mediation, the creatio
n of a link that did not exist before and
that to some degree modifies the original [goals].»

It is neither humans
nor non
humans that govern the situation. Rather the responsibility for the
action is distributed among the various actants.

: Ac
tion is seen as a composition of diverse subprograms.
When we, for example, think of ‹driving› then it depends on an
association of driver, car, petrol, street, environmental legislation, traffic
regulations, etc. «Action is simply not a property of humans

but of an
association of actants
, and this is the second meaning of technical
mediation. Provisional ‹actorial› roles may be attributed to actants only
because actants are in the process of exchanging competences, offering
one another new possibilities, n
ew goals, new functions.»

Reversible Blackboxing
: Blackboxing is a procedure in folding time and
space. It blacks out the development of events produced by the interplay
of both actors and artefacts. The common and mutual efforts of the actants
in a so called black box. A computer is a typical example of
such a black box. It is just seen as a ‹mute intermediary› that fulfils its
planned function. We do not think of what is happening inside this
complex machine unless it fails. But in case of fail
ure we become aware
of that tightly packed constellation. As long as an artefact functions, the
actants involved in the technical device remain invisible. Only when there
are unusual occurrences do we suddenly encounter a new assembly of
actants without kn
owing their performance. To get an idea of all the pre
conditions that must be met to act just look around and try to figure it out.
The number of actants varies depending on the situation and the point of
view from which an action takes place. Put in othe
r words: «When a
machine runs efficiently, when a matter of fact is settled, one need focus
only on its inputs and outputs and not its internal complexity. Thus,
paradoxically, the more science and technology succeed, the more opaque
and obscure they becom

: Delegation means that matter matters. States of affairs cannot
only be grasped by the observation of goals and functions. Programs of
action are also formed by the matter of expression. It makes a difference if
you are forced to do somethi
ng by physical conditions or by moral
prescriptions. So for example a speed bump is articulated as a translation
of the expression of slowing down speed so as not to put life at risk.
Delegation means an action type that bypasses the problematic dichotomy
of fact and belief. «In delegation it is not, as in fiction, that I am here and


Ibid., p 179.


Ibid., p 182 (italics in th
e original).


Ibid., p 304
. Perhaps we could draw a parallel with Heidegger's terms of ‹ready
hand› (Zuhandenheit)
and ‹present
hand› (Vorhandenheit). In contrast to Heidegger's conception we must keep in mind that
theory is based on t
he hypothesis that there is a symmetrical relationship between human
beings and technical artefacts.


elsewhere, that I am myself and someone else, but that an action, long
past, of an actor, long disappeared, is still active here, today, on me. I live
in the midst of technical

; I am folded into nonhumans.»

So the
conclusion is that human beings are not alone but inevitably worked into a
web of associated actants.

Agent technology

One can hardly find a concise definition of the term ‹agent›, even though it is
such a
widely used term. In this paper agent is a synonym for actant. Agents
can be described as delegates, as they are above. Agents as well as actants are
producers of effects and play their role as part of a design branch called
‹distributed intelligence› or s
omething similar. Due to developments in
information technology human beings get artificial mates that help us to
manage the tasks at issue. It is interesting that agents are described as
autonomous delegates that are on the team so to speak. The very shor
t joint
venture between me, my computer, the internet protocol, websites, the google
agents, etc. got the following definition of agent: «Agents are considered to be
autonomous (i.e., independent, not
controllable), reactive (i.e., responding to
events), p
active (i.e., initiating actions of their own volition), and social
(i.e., communicative). Sometimes a stronger notion is added (beliefs, desired,
intentions) realising intention notions for agents. Agents vary in their abilities;
e.g. they can be stati
c or mobile, or may or may not be intelligent. Each agent
may have its own task and/or role. Agents, and multi
agent systems are used as
a metaphor to model complex distributed processes.»

Furthermore there are endeavours to enhance the sociological acti
on concept.
In the sociology of technology the question is raised whether machines are
able to act. In this field of research you can find the concept of a gradually
differentiated agency of artefacts.

The proposed scale of agency encloses the
range from
passive to transactive.

The players are not the game

So far I have presented the actants as the players. But we must not forget the
game. Therefore I will now take a brief look at the conception of game that
best suits my purpose, in order to frame some
rudiments of an interface theory.
The game is often presented as the primordial phenomenon of living together
and so I will call it the

‹game of life›.

On the one hand the ‹game of life› is a scientific simulation that renders real
life processes. It is a

model of evolution based upon a small set of rules. These
rules are also called ‹genetic laws› and decide births, deaths and survivals,
represented as cells in a cellular automaton. In this context it is not necessary


Ibid., p 189 (italics in the original).



(access date: 2004


See for example
Rammert, Werner; Schulz
Schaeffer, Ingo (Hrsg.) (2002): Können Maschinen handeln?
Soziologische Beiträge zum Verhältnis von Mensch und Technik, Frankfurt
or browse the website of

‹Sozionik› (http://www.fernuni


Concerning this see
Rammert, Werner (2003): Technik in Aktion: Verteiltes Handeln in soziotechnischen
Konstellationen, S. 289
315 in: Thomas Christaller; Josef Wehner (Hrsg.), Autonome Maschinen,


to give a full description of the gam

I would just like to highlight two
points that are characteristic of games. First, one cannot foresee the course of
the game. And second, the course of the game is determined by the interaction
of the players. There is always an interdependence of the
players' behaviour.

On the other hand the ‹game of life› is a metaphor of ‹real life›. How could we
tell them apart? In any case, it stands for existence. And this means among
other things that the rules itself are subject to the processes of evolution. B
then, what is the game about? In David Cronenberg's film


a similar
question was asked: «What precisely is the goal of the game we are playing
now?» And the answer was as simple as: «You have to play the game to find
out why you're playing.» W
hat could I add…

The ‹game of life›, if I may simplify, is about being through action. My point
is that the very game is conceptualized as a chain or a network of actions. And
an action is, according to my proposition, the momentary association of
. Also, in order to come into play we just need an activated transition of
events. There is no single action but a succession of actions. Each action refers
to another action. And as long as an action takes place the game continues. So
the single units of
games are actions. Actions again constitute themselves by
attracting and binding actants to a temporal position which is related to pre
positions and post
positions. In this sense actants are
d by the game and
the game is formed by a continuation of

actants associating themselves.

And so there are various ways of playing games, just as there are various ways
of shuffling and dealing cards. The trajectory of the game is always
indeterminate because there are no possibilities of deciding the course of

game once and for all. The ‹game of life› is full of uncertainties because
actions arise from selections. And selections may or may not happen. In other
words, actions are contingent events. They are neither necessary nor
impossible, but placed in the

realm in between. Actions are reproduced by
constantly renewing themselves in different and unforeseen ways. Thus the
assumption of plain rationality is rather a figment of the imagination than a
sheer fact. Nevertheless actions are dependent on certain r
ules and certainly
count on them for the most part

although maybe as a constitutional illusion.
But the rules themselves are more or less in flux. Just think of Wittgenstein's
‹language game›. According to this, by playing games we are not solving
ms but endlessly tinkering with problems. However this is surely not a
meaningless process because further actions are generated and the ‹game of
life› continues. One could argue that this outline elicits a rather unsettling
picture. Be that as it may, it
should be taken into account that accidence is at
least to be considered as the opposite side of normalized regularity.

Observing action

In conclusion I will briefly outline my proposition of design (theory) as a
mode of observing (emerging) actions. I in
troduced actants in detail as the
constitutive elements of action. Their inter
action or co
action determines
action as basic events of the game. Thus action is considered as the main key


If yo
u are interested in this topic see for instance:


just browse the
net using the search words ‹John Conway's game of life



to design and the conception of interface. By observing actions desi
gn would
be put onto the right track to understand the interface. In other words,
«interface design should concern itself with
representing whole actions with
multiple agents


Interface is produced from the space between actions, from
the ‹inter
› of the

action. Its spatiotemporal texture emerges from ‹face to face›
relations among actants. By conceptualising interface as a fluctuating form of
instantly coupling multifarious actants the choreography of self
actions explains the story of the game

of life. In such a conception there is no
longer a user dealing with a preconfigured functional device aiming at a goal
but an actantial drift of actions. In this sense action could be described as a
kind of magic ‹bricolage›, an ongoing improvisation wit
hout knowing what is
coming next. Is that not the way wicked or ill
defined problems are treated just
to claim a planned action in hindsight? Rather than think of action as a rational
outcome of human beings it is time to observe action as mutual control o
actants. From this point of view we

but who is we?

are able to observe
actants facing actants and actions obs
erving actions. The question is

how and
when the joker

comes into the play.


Well then, let the actants play the game

and observe w
hat is going on

Biographical Note

Victor Zwimpfer, studied sociology and philosophy at the University of Basel.
Long experience in
software design
. Since 2001

at the university of art and design
(FHA Gestalt
ung und
Kunst). Major areas of research: interface theory, inter
action, virtual reality.


Victor Zwimpfer

Fachhochschule Nordwestschweiz

FHA Gestaltung und Kunst

Bahnhofstrasse 102


5000 Aarau

Mail: v.zwimpfer@fh


Laurel, Brenda (1993): Computers as Thea
tre, Boston, p 7 (italics in the original).


The concept of the joker as a wild card is portrayed in
Serres, Michel (1981): Der Parasit, Frankfurt
, p

See also
Baecker, Dirk (1993): Die Form des Unternehmens, Frankfurt
, p 51ff